| Abstract [eng] |
The examination of geometric max (min) stability in climatology, finance or insurance areas isn’t enough to examine one random variable, sometimes necessary to consider the whole system of random variables. In this master’s work our purpose is graduate from the geometric max (min) stability univariate case to the bivariate case. Extension random bivariate vectors can be made to the multivariate vectors. Research of bivariate distribution (Pareto, logistic) function submitted unexpected results. The examinations of distributions whose components are independent have not received geometric max (min) stability and have not received asymptotic max (min) stability. If bivariate distributions, whose vectors components are dependent are geometric min stable they do not necessarily will be geometric max stable and vice versa. Marginal distribution functions are geometric max (min) stable. |
